Siri to be more efficient? Apple claims its ReALM is better than OpenAI's GPT-4 at this task. What is it

ReALM could improve Siri's ability to understand context in a conversation, process onscreen content better as well as detect background activities.

Apple's AI researchers shed light on the company's AI plans for Siri in a research paper published this week. The paper introduced Reference Resolution As Language Modeling (ReALM) which is a conversational AI system aimed at improving reference resolution.

How does ReALM work?

The Apple logo is illuminated at a store in Munich, Germany. The researchers compared ReALM models against OpenAI's LLMs GPT-3.5 and GPT-4. (AP)

ReALM could improve Siri's ability to understand context in a conversation, process onscreen content better as well as detect background activities.

The system can also convert conversational, onscreen, and background processes into a text format which can then be processed by large language models (LLMs).

What researchers said on ReALM compared to OpenAI models?

The researchers compared ReALM models against OpenAI's LLMs GPT-3.5 and GPT-4 which power the free ChatGPT and the paid ChatGPT Plus. In the paper, the researchers said, "We demonstrate large improvements over an existing system with similar functionality across different types of references, with our smallest model obtaining absolute gains of over 5% for onscreen references. We also benchmark against GPT-3.5 and GPT-4, with our smallest model achieving performance comparable to that of GPT-4, and our larger models substantially outperforming it.”

How ReALM performed in comparison to GPT-4?

Apple's annual Worldwide Developers Conference (WWDC) is scheduled for June 10 in which AI could be the focus as per Greg Joswiak, Apple's senior vice president of global marketing.